By: Dummond D. Slow (mental.delete@this.protozoa.us), November 18, 2020 4:27 pm
Room: Moderated Discussions
Doug S (foo.delete@this.bar.bar) on November 18, 2020 3:02 pm wrote:
> Dummond D. Slow (mental.delete@this.protozoa.us) on November 18, 2020 7:58 am wrote:
> > Doug S (foo.delete@this.bar.bar) on November 18, 2020 7:44 am wrote:
> > >
> > > If so, then the only argument is that 4 big cores WITH SMT along with 4 little cores is faster
> > > than 4 big cores without SMT and 4 little cores. Which is obvious, but then you could add 4
> > > more little cores, or one more big core, to reach that level of performance without all the
> > > extra design/verification/security headaches of SMT. Before someone says "but that costs more
> > > because the die is bigger", yes, but we're talking about increasing the M1's die size maybe
> > > 3% - and for something that will benefit ALL MT loads, not just the ones are helped by SMT.
> >
> > Oh, the die size argument. Well if adding a core costs so little, why didn't Apple already do it?
>
>
> Because this is the CPU designed for the lowest end entry level Macs. And incidentally
> almost certainly the exact same piece of silicon that will be called A14X when
> it is installed in the next generation iPad Pro early next year.
>
> There will be Apple Silicon CPUs with more cores for higher end Macs, all the way up to the Mac Pro.
> I'm sure no matter how fast the Mac Pro is, someone will make the argument "it could have been even
> faster if the big cores had SMT". I suppose whether that question matters depends on whether it ends
> up clearly faster than any x86 workstation available at any price at the time it is released.
My point was that you that it is problematic to use "this costs so little silicon so it is almost free" or "apple doesn't have to care about silicon budget" as a rebuttal to something when it is clear that at least to some extent they still do care about sizes of stuff - the chip clearly could have been bigger but is not.
This same fallacy really rises up with every company. When you look at for example AMD chips, the cores are suprisingly little. "What, 4 cores only take 20% of the die or less, why didn't they add 2x as much, I would pay extra for that! Isn't that dumb?" Or, "what, those 4MB of L3 cache they cut was like just 4 mm! WHY YOU IDIOTS?"
Companies clearly have reasons to do stuff like that.
> Dummond D. Slow (mental.delete@this.protozoa.us) on November 18, 2020 7:58 am wrote:
> > Doug S (foo.delete@this.bar.bar) on November 18, 2020 7:44 am wrote:
> > >
> > > If so, then the only argument is that 4 big cores WITH SMT along with 4 little cores is faster
> > > than 4 big cores without SMT and 4 little cores. Which is obvious, but then you could add 4
> > > more little cores, or one more big core, to reach that level of performance without all the
> > > extra design/verification/security headaches of SMT. Before someone says "but that costs more
> > > because the die is bigger", yes, but we're talking about increasing the M1's die size maybe
> > > 3% - and for something that will benefit ALL MT loads, not just the ones are helped by SMT.
> >
> > Oh, the die size argument. Well if adding a core costs so little, why didn't Apple already do it?
>
>
> Because this is the CPU designed for the lowest end entry level Macs. And incidentally
> almost certainly the exact same piece of silicon that will be called A14X when
> it is installed in the next generation iPad Pro early next year.
>
> There will be Apple Silicon CPUs with more cores for higher end Macs, all the way up to the Mac Pro.
> I'm sure no matter how fast the Mac Pro is, someone will make the argument "it could have been even
> faster if the big cores had SMT". I suppose whether that question matters depends on whether it ends
> up clearly faster than any x86 workstation available at any price at the time it is released.
My point was that you that it is problematic to use "this costs so little silicon so it is almost free" or "apple doesn't have to care about silicon budget" as a rebuttal to something when it is clear that at least to some extent they still do care about sizes of stuff - the chip clearly could have been bigger but is not.
This same fallacy really rises up with every company. When you look at for example AMD chips, the cores are suprisingly little. "What, 4 cores only take 20% of the die or less, why didn't they add 2x as much, I would pay extra for that! Isn't that dumb?" Or, "what, those 4MB of L3 cache they cut was like just 4 mm! WHY YOU IDIOTS?"
Companies clearly have reasons to do stuff like that.
